On Thu, Jun 25, 2009 at 03:39:47PM -0500, Shawn Walker wrote:
> [email protected] wrote:
>> On Tue, Jun 23, 2009 at 11:13:18PM -0500, Shawn Walker wrote:
>>> One final concern I had is that I noticed that touch_manifest now   
>>> performs a GET instead of a HEAD (I checked the depot server logs).  
>>> Was this a limitation of pycurl?  The 'HEAD' operations should be 
>>> faster for the server...
>>
>> It shouldn't be doing this.  You may have found a bug in PyCurl/libcurl.
>> In cases where we perform a "HEAD" I tell the handle to set
>> CURLOPT_NOBDOY.  According to the libcurl documentation, this should
>> make the client perform a HEAD.  If it's not doing this, then we have a
>> bug.  I'll investigate this further.
>
> I believe this is due to:
>
> modules/client/transport/engine.py:
>   line 505: this should be an elif instead of 'if' because, based on my  
> reading of the libcurl docs [1], HEAD should only use pycurl.NOBODY, and  
> using pycurl.HTTPGET will switch to a GET.

Yes, that's right.  That was supposed to be an if, elif, else block.
I've been staring at this code for too long, I guess.

-j

_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss

Reply via email to